This MCP server provides two AI-callable tools for working with currencies: convert_currency and list_currencies. The convert_currency tool takes an amount, a source currency code, and a target currency code to return the converted value. The list_currencies tool fetches and returns the supported currencies from an external Currency API. You can test and interact with these tools through standard MCP JSON-RPC requests, or by using the included test_mcp.py script to see the available tools and their expected payloads. When integrated with Claude Desktop, you can connect to the server by configuring the mcpServers.currency entry in claude_desktop_config.json, pointing to the UV runner that executes currency.py and optionally providing your Currency API key via the CURRENCY_API_KEY environment variable. This setup enables Claude to call the two tools directly from your local MCP server.

Prerequisites:

• Python 3.10+
• UV installed (for dependency management and running the MCP server)
• Access to a Currency API key

Installation steps:

1. Clone the repository: git clone https://github.com/manishh/mcp-server-demo cd <project-directory>

2. Create and activate a Python virtual environment (optional but recommended):

   if using Python venv

   python3 -m venv .venv source .venv/bin/activate # macOS/Linux .venv\Scripts\activate # Windows

3. Install dependencies and set up the environment: uv sync

4. Export your Currency API key (required by the server): export CURRENCY_API_KEY=your_api_key_here # macOS/Linux set CURRENCY_API_KEY=your_api_key_here # Windows

5. Start the MCP server: uv run currency.py

6. (Optional) Test available tools locally: python test_mcp.py

Notes:

• Ensure the CURRENCY_API_KEY environment variable is set before starting the server.
• The server uses stdio transport for integration with local AI agents like Claude Desktop.

Tips and troubleshooting:

• If you encounter authentication or key issues, verify that CURRENCY_API_KEY is correctly exported and accessible to the process running currency.py.
• For Claude Desktop integration, ensure the absolute path to the UV executable and the project directory are correctly configured in claude_desktop_config.json as shown in the README example.
• The server is educational and uses an external Currency API; rate limits or API changes may affect tool responses.
• The two available tools are:
  • convert_currency: converts a given amount from one currency to another.
  • list_currencies: returns a list of supported currencies from the external API.
• If you modify currency.py or related files, re-test with python test_mcp.py to verify tool availability and RPC formats.